Based On Multitask Cascaded Convolutional Networks MTCNN, Implementation of Behavioral Drowsiness Detection System

Tiredness and driver drowsiness rank among the most common reasons for car accidents. The number of accidents and deaths substantially increases each year for a variety of reasons. In order to prevent traffic accidents and consequently preserve lives and public property, structures that warn drivers are necessary. The Driver Drowsiness Detection System (DDDS) was developed and put into use in this research to protect the driver from accidents. This study sought to create a low-cost DDDS that could quickly alert a motorist to ensure their safety, as well as a DDDS that utilized a high-resolution camera and Multi-Cascaded Convolutional Neural Networks to detect driver sleepiness with high accuracy (MTCNN). DDDS is recommended for identifying driver intoxication. The method is based on the driver's behaviour utilizing visual cues like closed or open eye movement recorded by a high-resolution camera. Eyes Landmarks are examined at the direction of measuring a car user's fatigue in a present frame of reference, leading to the discovery of a novel restriction called "Eyes Aspect Ratio." The drive casing alarm was then initiated using the MTCNN output. The experiment employed an image with a frame rate of 60 frames per second (f/s) and a resolution of 4320 eyes. According to testing data, the accuracy of drowsiness detection was greater than 99.9 percent in good illumination and greater than 99.8 percent in bad lighting. The new study outperformed a number of other studies in terms of drowsiness detection accuracy.